EDB Postgres for Kubernetes 0.6 release notes v1
We are in the process of migrating the documentation from the previous EDB Postgres for Kubernetes name to the new EDB CloudNativePG Cluster name. You may see references to both names in this documentation.
Released: 4 Dec 2020
| Type | Description |
|---|---|
| Feature | Add Point-In-Time Recovery based on timestamp, target name, or transaction Id, as well as the specification of the timeline, through a new bootstrap method option called recoveryTarget. |
| Feature | Add Synchronous Streaming Replication support through the minSyncReplicas and maxSyncReplicas cluster options, defining respectively the expected minimum and maximum number of synchronous standby servers at any time (disabled by default) |
| Feature | Support EDB Postgres Advanced Server (EPAS). |
| Feature | Configure initdb options for the bootstrap of an empty cluster (initDb). |
| Feature | Enable/Disable Redwood compatibility level with EPAS. |
| Feature | Extend the instance manager with a new framework for the export of metrics for Prometheus - currently supporting pg_stat_archiver only. |
| Feature | Use Kubernetes jobs instead of init containers to perform cluster initialization procedures (including standby creation and recovery) and improve their observability. |
| Feature | Record Kubernetes events to be used by kubectl describe and kubectl get events. |
| Feature | Introduce Kubernetes expectations for Pods, PVCs, and Jobs to prevent race conditions. |
| Feature | Set application_name in PostgreSQL to the name of the Pod/instance. |
| Feature | The fullRecovery bootstrap mode has been renamed to recovery to address also Point-In-Time Recovery |
| Fix | Bug fixes and code improvements. |